By Heidi Ledford of Nature magazine

When the US Department of Agriculture (USDA) announced this month that it did not have the authority to oversee a new variety of genetically modified (GM) Kentucky bluegrass, it exposed a serious weakness in the regulations governing GM crops. These are based not on a plant's GM nature but on the techniques used for its genetic modification. With changing technologies, the department says that it lacks the authority to regulate newly created transgenic crops.

The grass, a GM variety of Poa pratensis, is still in the early stages of development by Scotts Miracle-Gro, a lawn-care company based in Marysville, Ohio. The grass has been genetically altered to tolerate the herbicide glyphosate, which would make it easier to keep a lawn weed-free. On 1 July, secretary of agriculture Tom Vilsack wrote to the company to say that the variety "is not subject" to the same regulations that govern other GM crops. The decision allows Scotts to bypass the years of environmental testing and consultation typically required by the regulators for GM plants, although the company says there are no plans to market this particular variety.

The grass can evade control because the regulations for GM plants derive from the Federal Plant Pest Act, a decades-old law intended to safeguard against plant pathogens from overseas. Previous types of GM plants are covered because they were made using plant pathogens. The bacterium Agrobacterium tumefaciens--which can cause tumors on plants--shuttled foreign genes into plant genomes. Developers then used genetic control elements derived from pathogenic plant viruses such as the cauliflower mosaic virus to switch on the genes.

By revealing similar elements in plants' DNA, genome sequencing has liberated developers from having to borrow the viral sequences. And Agrobacterium is not essential either; foreign genes can be fired into plant cells on metal particles shot from a 'gene gun'. Scotts took advantage of both techniques to construct the herbicide-resistant Kentucky bluegrass that put the USDA's regulatory powers to the test.

"The Plant Pest Act was completely inappropriate for regulating biotech crops, but the USDA jury-rigged it," says Bill Freese, science-policy analyst at the Center for Food Safety in Washington DC. "Now we can foresee this loophole getting wider and wider as companies turn more to plants and away from bacteria and other plant-pest organisms." The USDA has not made public any plans to close the loophole and has also indicated that it will not broaden its definition of noxious weeds, a class of plants that falls under its regulatory purview, to facilitate the regulation of GM crops.

Nevertheless, Agrobacterium is still industry's tool of choice for shuttling in foreign genes, says Johan Botterman, head of product research at Bayer BioScience in Ghent, Belgium. The technique is well established for many crops, and particle bombardment is less predictable, often yielding multiple, fragmented insertions of the new gene.

But Agrobacterium isn't suitable for some new techniques. Many companies are developing 'mini-chromosomes' that can function in a plant cell without needing to be integrated into the plant's genome. Last summer, agribusiness giant Syngenta, based in Basel, Switzerland, conducted the first field trials of maize (corn) containing engineered mini-chromosomes, and showed that the mini-chromosomes, which carried multiple genes for insect and herbicide resistance, were stable in the field. "I would expect that by the end of the decade, this technology will be well used by many as a way to deliver large stacks of genes to plants," says Roger Kemble, head of technology scouting for Syngenta.

Other techniques under development insert foreign genes into designated sites in the genome, unlike the near-random scattering generated by Agrobacterium. In 2009, researchers at Dow AgroSciences in Indianapolis, Indiana, and Sangamo BioSciences in Richmond, California, announced that they had used enzymes called zinc-finger nucleases to insert a gene for herbicide resistance at a specific site in the maize genome (V. K. Shukla et al. Nature 459, 437-441; 2009). Bayer is interested in harnessing other enzymes called 'meganucleases' to do the same type of targeted engineering, a strategy that Botterman says may make it possible to introduce multiple new traits into existing GM crops.

Regulators need to adapt to these new techniques, or run the risk of over- or under-regulating GM plants, says Roger Beachy, a plant biologist at Washington University in St Louis, Missouri, and former head of the USDA's National Institute for Food and Agriculture. The Kentucky bluegrass decision drives this point home, he says: "It really speaks to the importance of reviewing the regulatory process periodically to ensure that it is keeping up with the advances in technology."

This article is reproduced with permission from the magazine Nature. The article was first published on July 20, 2011.